At 07:26 PM 7/12/2008, Terry Bouricius wrote:
By raising one-sided objections to any particular reform proposal that is being seriously considered, the net effect is most likely to be to shore up the status quo, rather than to advance one's favored method. If election method experts put their united effort into explaining why current plurality voting is bad, rather than attacking other election reformers' efforts, we would all be better off. That is not to say, of course, that we shouldn't continue this thoughtful behind the scenes discussion about pros and cons of various methods.
The problem is that IRV is being proposed for nonpartisan elections, where it is *not* an improvement. From what I now know, I'm directly opposed to IRV as a replacement for top-two runoff in nonpartisan elections, where it clearly produces worse results. And comparing it to plurality, in nonpartisan elections, it produces the *same* results.
It is only in partisan elections that IRV has shown an ability to, sometimes, improve results over Plurality.
In other words, in nonpartisan elections, I do not consider IRV a "reform," it is a mistake, at best useless, extra work for no gain. Where a true majority is required, and if such is no longer required when IRV is implemented, it is a step backward.
If the desire is to avoid *some* runoffs, Bucklin will do a better job, for less cost and many of the same benefits. Avoiding *all* runoffs is, again, a step away from democracy, and it harms results.
IRV, as I mentioned, from what experience we now have in the U.S. always, in these nonpartisan elections (it hasn't been used for partisan elections yet, in this modern spurt of implementations), is always electing the plurality winner from the first round. (And the preference order is more deeply preserved than that, the runner=up in the first round is remaining the runner-up after transfers, in the nine "instant runoffs" so far). In real runoffs, the runner-up in the first round wins about a third of the time, with a "comeback election." FairVote showed 29% comebacks in a series of federal primary elections in Texas.
Why was IRV used in the U.S. at one time, for primary elections, and discontinued? I understand that the reason was that the vote transfers were not changing results, so this really isn't anything new. The Australians also know this about IRV, that, even in partisan elections, vote transfers tend to favor the plurality leader. And, of course, with Optional Preferential Voting, used in Queensland and NSW, there are a high percentage of exhausted ballots, and apparently that percentage is growing. FairVote points to Australia as an example of the use of "IRV," but in the U.S., "IRV" is always optional preferential voting, OPV, a different animal than standard, Australian, fully-ranked PV (not to mention that here, the ranking is typically limited to three ranks, thus guaranteeing more exhausted ballots if voters vote sincerely in an election with many candidates, as in San Francisco where there may be over twenty candidates on the ballot.
No, Terry, for Fort Collins, from what I can see, IRV is a Bad Idea, and it is being promoted without regard to the actual needs of that community. They are only pawns in a game. Jan Kok actually lives there. He's supposed to set aside his knowledge about this? Why? Political strategy? He already knows that it's not going to buy him anything if he refrains. Every IRV implementation so far in the U.S. has been a mistake, and, apparently, hardly anybody is actually taking a careful look at the results. Especially not FairVote, and especially not you, Mr. Bouricius. Using IRV in partisan elections is a more arguable case. It's known to improve results, sometimes, and, there, the question is more subtle, i.e., perhaps there are better and cheaper reforms. But with nonpartisan elections, no. Don't like paying for runoff elections? Don't fall for the false promise of IRV, it usually won't find a majority, it merely pretends to. (If you don't think so, when, then, in a jurisdiction that requires a majority, why take that provision out of the law? Why not just leave it in and then use IRV to find it? But you know what would happen.)
To complete the idea, don't like paying for runoffs? Eliminate the majority requirement. That is exactly what is being done by implementing IRV, and the results stay the same. But want to *reduce* runoffs? You can use IRV, it's true, and this is what Robert's Rules is talking about. But even there, there are much simpler methods; Bucklin is not only cheaper to count, but it also is probably more efficient at finding a majority of votes, since it ends up counting all of them. Bucklin with a true majority required is a quite sophisticated method, for all its simplicity. IRV with a true majority required is much better than "election by Plurality" -- which includes IRV when it elects with a mere plurality of the votes, as you would have realized if you paid closer attention to all that writing I did in Talk on Wikipedia. That's what Robert's Rules is saying!
But, of course, if a majority is required or there is a runoff or other process, then IRV no longer satisfies this offensive Later No Harm criterion, which is why the ballot instruction you wrote into the proposed legislation in Vermont you authored was incorrect. Later No Harm is incompatible with a majority requirement. Which means that Later No Harm is incompatible with *democracy*.
---- Election-Methods mailing list - see http://electorama.com/em for list info
